U.S. patent number 6,359,288 [Application Number 09/064,242] was granted by the patent office on 2002-03-19 for nanowire arrays.
This patent grant is currently assigned to Massachusetts Institute of Technology. Invention is credited to Mildred S. Dresselhaus, Jackie Y. Ying, Lei Zhang, Zhibo Zhang.
United States Patent |
6,359,288 |
Ying , et al. |
March 19, 2002 |
**Please see images for:
( Certificate of Correction ) ** |
Nanowire arrays
Abstract
An array of nanowires having a relatively constant diameter and
techniques and apparatus for fabrication thereof are described. In
one embodiment, a technique for melting a material under vacuum and
followed by pressure injection of the molten material into the
pores of a porous substrate produces continuous nanowires. In
another embodiment, a technique to systematically change the
channel diameter and channel packing density of an anodic alumina
substrate includes the steps of anodizing an aluminum substrate
with an electrolyte to provide an anodic aluminum oxide film having
a pore with a wall surface composition which is different than
aluminum oxide and etching the pore wall surface with an acid to
affect at least one of the surface properties of the pore wall and
the pore wall composition.
Inventors: |
Ying; Jackie Y. (Winchester,
MA), Zhang; Zhibo (Cambridge, MA), Zhang; Lei
(Cambridge, MA), Dresselhaus; Mildred S. (Arlington,
MA) |
Assignee: |
Massachusetts Institute of
Technology (Cambridge, MA)
|
Family
ID: |
21930038 |
Appl.
No.: |
09/064,242 |
Filed: |
April 22, 1998 |
Current U.S.
Class: |
257/14; 205/202;
257/28; 257/9; 257/E21.251; 428/472.2; 428/606; 428/614; 428/629;
977/762 |
Current CPC
Class: |
B82B
1/00 (20130101); B82Y 30/00 (20130101); C30B
11/00 (20130101); H01L 21/31111 (20130101); H01L
21/76885 (20130101); C30B 11/00 (20130101); C30B
29/02 (20130101); C30B 11/00 (20130101); C30B
29/605 (20130101); C30B 29/605 (20130101); C30B
29/02 (20130101); Y10S 977/888 (20130101); Y10S
977/762 (20130101); Y10S 977/781 (20130101); Y10T
428/1259 (20150115); Y10T 428/12486 (20150115); Y10T
428/12431 (20150115) |
Current International
Class: |
C30B
11/00 (20060101); H01L 21/02 (20060101); H01L
21/311 (20060101); H01L 029/15 () |
Field of
Search: |
;257/9,28,30,29,14
;205/173,174,202 ;428/606,613,614,620,629,472.2 ;438/962 |
References Cited
[Referenced By]
U.S. Patent Documents
Foreign Patent Documents
Other References
"Nanowire Arrays Composites," Huber et al., Science, vol. 263, Feb.
11, 1994.* .
Routkevitch et al., Nonlithographic Nan-Wire Arrays: Fabrication,
Physics, and Device Applications, IEEE Transactions on Electron
Devices, vol. 43, No. 10, Oct. 1996.* .
C.A. Huber, et al. "Nanowire Array Composites," Science vol. 263,
pp. 800-802, Feb. 11, 1994. .
Dmitri Routkevitch, et al. "Nonlithographic Nano-Wire Arrays:
Fabrication, Physics, and Device Application," IEEE Transactions on
Electron Devices, vol. 43, No. 10, pp. 1646-1658, Oct., 1996. .
Dmitri Routkevitch, et al. "Porous Anodic Alumina Templates for
Advanced Nanofabrication," Electrochemical Society Proceedings,
vol. 97-7, pp. 350357. .
H. M. Cheng, *F. Li, and G. Su--Institute of Metal Research,
Academia Sinica, 72 Wenhua Rd., Shenyang, 110015 China H.Y. Pan and
L. L. He--Lab for Atomic Imaging of Solids, Institute of Metal
Research, Academia Sinica, 72 Wenhua Rd., Shenyang, 110015 China X.
Sun--Department of Physics, Massachusetts Institute of Technology,
Cambridge Massachusetts 02139 M.S. Dresselhaus--Department of
Electrical Engineering and Computer Scients and Department of
Physics, Massachusetts Institute of Technology, Cambridge, MA
02139--"Large-Scale and Low-Cost Synthesis of Single-Walled Carbon
Nanotubes by Catalytic Pyrolysis of Hydrocarbons"--pp. 1-3, (Jan.
13, 1998). *Also at Department of Physics, Massachusetts Institute
of Technology, Cambridge, MA 02139. .
"A New Family of Mesoporous Molecular Sieves Prepared with Liquid
Crystal Templates", Beck et al., J. Am. Chem. Soc. 1992, 114,
10834-10843. .
"Fabrication of quantum wire structures by in-situ gas etching and
selective-area metalorganic vapor phase epitaxy regrowth",
Shimoyama et al., Journal of Crystal Growth 145 (1994) 734-739.
.
"Fabrication of Quantum Wire and Minute Buried Heterostructure by
In Situ Etching and Selective MOCVD Growth", Ogura, et al., Part I,
No. 2B, Feb. 1990, pp. 1353-1356 and last page unnumbered. .
"Structural Features of Oxide Coatings on Aluminum", Keller, et
al., Journal of the Electrochemical Society, vol. 100, No. 9, Sep.
1953, pp. 411-419. .
"Properties of Porous Anodic Aluminum Oxide Films as Membranes",
Itaya, et al., Journal of Chemical Engineering of Japan, vol. 17,
No. 5 1984, pp. 514-520. .
"The formation of controlled-porosity membranes from anodically
oxidized aluminum", Furneaux, et al., Nature vol., 337, Jan. 12,
1989. .
"Micropolarizer made of the anodized alumina film", Saito, et al.,
Appl. Phys. Lett. 55 (7), Aug. 14, 1989. .
"A Composite Palladium and Porous Aluminum Oxide Membrane for
Hydrogen Gas Separation", Konno, et al., Journal of Membrane
Science, 37, (1988) 193-197. .
"Electrochemical Fabrication of CdS Nanowire Arrays in Porous
Anodic Aluminum Oxide Templates", Routkevitch, et al., J. Phys.
Chem. 1996, 100, 14037-14047. .
"Ordered Metal Nanohole Arrays Made by a Two-Step Replication of
Honeycomb Structures of Anodic Alumina", Masuda, et al., Science,
vol. 268, Jun. 9, 1995. .
"Preparation of Ultrafine Carbon Tubes in Nanochannels of an Anodic
Aluminum Oxide Film", Kyotani, et al., 1996, American Chemical
Society, Chem. Mater., vol. 8, No. 8. .
"Injection Under High Pressure of Molten Metals into Crystalline
Materials Containing Regular Cavities", Bogomolov, Soviet
Physics--Solid State, vol. 13, No. 3, Sep., 1971. .
"Resistivity Anomaly in Thin Bi Wires: Possibility of a
One-Dimensional Quantum Size Effect", Gurvitch, Journal of Law
Temperature Physics, vol. 38, Nos. 5/6, 1980. .
"Nanowire Array Composites", Huber, et al., Science, vol. 263, Feb.
11, 1994. .
"The Physical Review, The Dynamics of Capillary Flow", Washburn,
Second Series, Mar., 1921, vol. XVII, No. 3. .
"Wetting: statics and dynamics", P.G. de Gennes, Reviews of Modern
Physics, vol. 57, No. 3, Part I, Jul. 1985. .
"Capillarity-induced filling of carbon nanotubes", Ajayan, et al.
Nature, vol. 361, Jan. 28, 1993. .
"Nucleation and growth of porous anodic films on aluminum",
Lanzerotti, et al., Nature vol. 272, Mar. 30, 1978. .
"Computer Simulation of Inorganic Membrane Morphology Part 3.
Anodic Alumina Films and Membranes", Randon, et al., Journal of
Colloid and Interface Science 169, 335-341 (1995). .
"The morphology and mechanism of formation of porous anodic films
on aluminum", O'Sullivan et al., Proc. Roy. Soc. Lond. A. 317,
511-543 (1970). .
"Variation of the connectivity of the electron constant-energy
surface in Bi under pressure", Brandt, et al., Sov. Phys. JETP
47(5), May 1978. .
"Nanocapillarity and Chemistry in Carbon Nanotubes", Ugarte, et
al., Science, vol. 274, Dec. 13, 1996..
|
Primary Examiner: Jackson, Jr.; Jerome
Assistant Examiner: Baumeister; Bradley Wm.
Attorney, Agent or Firm: Daly, Crowley & Mofford,
LLP
Government Interests
GOVERNMENT RIGHTS
This invention was made with Government support under Contract No.
CTS-9257223, awarded by the National Science Foundation and
Contract No. N00167-92-K-0052 awarded by the Department of the
Navy. The Government has certain rights in this invention.
Parent Case Text
RELATED APPLICATIONS
This application claims priority under 35 U.S.C. .sctn.119(e) from
Provisional Application No. 60/044,008, filed Apr. 24, 1997.
Claims
What is claimed is:
1. An nanowire array comprising:
a substrate having a plurality of non-interconnected pores each of
the plurality of pores having a mean pore diameter which does not
vary by more than 100% along the length of the pore; and
a material continuously filled in each of the plurality of pores of
the substrate such that the material in each of the plurality of
non-interconnected pores corresponds to a single crystal quantum
wire with each of the single crystal quantum wires having
substantially the same crystal orientation.
2. The nanowire array of claim 1 wherein:
each of the a single crystal quantum wires is provided having a
length not less than three microns; and
the diameters of each of the plurality of pores are in the range of
about 1 nanometer to about 300 nanometers.
3. The nanowire array of claim 2 wherein the diameters of each of
the plurality of pores are in the range of about 1 nanometer to
about 100 nanometers.
4. The nanowire array of claim 2 wherein the diameters of each of
the plurality of pores are in the range of about 1 nanometer to
about 50 nanometers.
5. The nanowire array of claim 2 wherein the mean diameter of each
of the pores do not vary by more than 50% along the length of the
pore.
6. The nanowire array of claim 2 wherein the mean diameter of each
of the pores do not vary by more than 20% along the length of the
pore.
7. The nanowire array of claim 1 wherein each of the plurality of
pores in said substrate have a pore wall surface exposed to an
electrolyte solution selected from the group of: (1) an H.sub.2
SO.sub.4 solution; (2) an solution of H.sub.2 C.sub.2 O.sub.4 ; and
(3) an H.sub.3 PO.sub.4 solution to thereby change the composition
of the pore walls to reduce a contact angle between the pore walls
and the material in the molten state.
8. The nanowire array of claim 1 wherein said material continuously
filling each of the plurality of pores is solidified from a
pressure injected metal in a molten state.
9. An array of nanowires comprising:
a substrate having a plurality of pores provided therein, with each
of the pores having a pore wall surface composition provided by
last exposing the pore walls to an H.sub.2 SO.sub.4 electrolyte
solution to provide the pore walls having a contact angle which
reduces the amount of pressure needed to fill the pore using a
pressure injection process; and
a nanowire disposed in each of the pores, each of the nanowires
having a single crystal structure and each of the single crystal
nanowires having substantially the same crystal orientation and a
packing density greater than 4.6.times.10.sup.10 cm.sup.-2.
10. The array of nanowires as recited in claim 9 wherein said means
for supporting comprises an anodic aluminum oxide film.
11. The array of nanowires as recited in claim 9 wherein said means
for supporting comprises a substrate comprised of a mesoporous
material.
12. The array of nanowires as recited in claim 11 wherein said
mesoporous material is mesoporous siliceous MCM-41.
13. The array of nanowires as recited in claim 9 wherein each of
the nanowires have an average wire diameter in the range of about 1
nm to about 20 nm.
Description
FIELD OF THE INVENTION
This invention relates generally to nanowires and more particularly
to nanowires having a diameter which is relatively small and
uniform and techniques for fabrication thereof.
BACKGROUND OF THE INVENTION
As is known in the art, a nanowire refers to a wire having a
diameter typically in the range of about one nanometer (nm) to
about 500 nm. Nanowires are typically fabricated from a metal or a
semiconductor material. When wires fabricated from metal or
semiconductor materials are provided in the nanometer size range,
some of the electronic and optical properties of the metal or
semiconductor materials are different than the same properties of
the same materials in larger sizes. Thus, in the nanometer-size
range of dimensions, the physical dimensions of the materials may
have a critical effect on the electronic and optical properties of
the material.
Quantum confinement refers to the restriction of the electronic
wave function to smaller and smaller regions of space within a
particle of material referred to as the resonance cavity.
Semiconductor structures in the nanometer size range exhibiting the
characteristics of quantum confinement are typically referred to as
zero-dimension (OD) quantum dots or more simply quantum dots when
the confinement is in three dimensions. Quantum dots are provided
from crystalline particles having a diameter less than about ten
nanometers which are embedded within or on the surface of an
organic or inorganic matrix and which exhibit quantum confinement
in three directions.
Similarly, when the confinement is in one dimension, the structures
are referred to as 2D quantum well superlattices or more simply
"quantum wells." Such superlattices are typically generated by the
epitaxial growth of multi-layer active crystals separated by
barrier layers. The 2D quantum wells have typically enhanced
carrier mobility and also have characteristics such as the quantum
Hall effect and quantum confined Stark effect. 2D quantum well
superlattice structures also typically have magnetoresistance and
thermoelectric characteristics which are enhanced relative to 3D
materials. One problem with quantum well superlattices, however is
that they are relatively expensive and difficult to produce and
fabrication of quantum well superlattices are limited to several
material systems including group IV semiconductors (such as SiGe),
group III-V compounds (such as GaAs), group II-VI compounds (such
as CdTe) and group IV-VI compounds (such as PbTe).
When the quantum confinement is in two dimensions, the structures
are typically referred to as a one-dimensional quantum wires or
more simply as quantum wires. A quantum wire thus refers to a wire
having a diameter sufficiently small to cause confinement of
electron gas to directions normal to the wire. Such two-dimensional
(2D) quantum confinement changes the wire's electronic energy
state. Thus, quantum wires have properties which are different from
their three-dimensional (3D) bulk counterparts.
For example, metallic wires having a diameter of 100 nm or less
have specific properties typically referred to as quantum
conduction phenomena. Quantum conduction phenomena include but are
not limited to: (a) survival of phase information of conduction
electrons and the obviousness of the electron wave interference
effect; (b) breaking of Ohm's Law and the dependence of the
electrical conductivity and thermal conductivity characteristics of
the wire on the configuration, diameter and length of the metal;
(c) greater fluctuation of wire conductivity; (d) noises observed
within the material depend upon the configuration of the sample and
the positions of impurity atoms; (e) a mark surface effect is
produced; and (f) visible light enters throughout the thin wire
causing a decrease in conductivity.
In transport-related applications, quantum wire systems exhibit a
quantum confinement characteristic which are enhanced relative to
quantum well systems. It is thus desirable to fabricate quantum
wire systems or more generally nanowire systems for use in
transport-related applications. One problem with nanowire systems,
however, is that it is relatively difficult to fabricate nanowires
having a relatively small, uniform diameter and a relatively long
length.
One technique for fabricating quantum wires utilizes a micro
lithographic process followed by metalorganic chemical vapor
deposition (MOCVD). This technique may be used to generate a single
quantum wire or a row of gallium arsenide (GaAs) quantum wires
embedded within a bulk aluminum arsenide (AlAs) substrate. One
problem with this technique, however, is that microlithographic
processes and MOCVD have been limited to GaAs and related
materials. It is relatively difficult to generate an array of
relatively closely spaced nanowires using conventional
microlithographic techniques due to limitations in the tolerances
and sizes of patterns which can be formed on masks and the MOCVD
processing required to deposit the material which forms the wire.
Moreover, it is desirable in some applications, to fabricate two
and three dimensional arrays of nanowires in which the spacing
between nanowires is relatively small.
Another problem with the lithographic-MOCVD technique is that this
technique does not allow the fabrication of quantum dots or quantum
wires having diameters in the 1-100 nanometer range. Moreover, this
technique does not result in a degree of size uniformity of the
wires suitable for practical applications.
Another approach to fabricate nanowire systems which overcomes some
of the problems of the lithographic technique, involves filling
naturally occurring arrays of nanochannels or pores in a substrate
with a material of interest. In this approach, the substrate is
used as a template. One problem the porous substrate approach is
that it is relatively difficult to generate relatively long
continuous wires having relatively small diameters. This is partly
because as the pore diameters become small, the pores tend to
branch and merge partly because of problems associated with filling
relatively long pores having relatively small diameters with a
desired material.
Anodic alumina and mesoporous materials, for example, each are
provided having arrays of pores. The pores can be filled with an
appropriate metal in a liquid state. The metal solidifies resulting
in metal rods filling the pores of the substrate. Surface layers of
the substrate surrounding the rods are then removed, by etching for
example, to expose the ends of the metal rods. In some applications
the rods can be chemically reacted to form semiconductor materials.
Some substrate materials, however, such as anodic alumina are not
suitable host templates for nanowires due to the lack of a
systematic technique to control pore packing density, pore diameter
and pore length in the anodic alumina.
Nevertheless, porous materials such as anodic alumina have been
used to synthesize a variety of metal and semiconductor
nanoparticles and nanowires by utilizing chemical or
electrochemical processes to fill pores in the anodic alumina. Such
liquid phase approaches, however, have been limited to Nickel (Ni),
Paladium (Pd), Cadmium sulfide (CdS) and possibly Gold (Au) and
Platinum (Pt). One problem with chemical or electrochemical
processes is that success of the processes depends upon finding
appropriate chemical precursors. Another problem with this approach
is that it is relatively difficult to continuously fill pores
having a relatively small diameter and a relatively long length
e.g. a length greater than 2.75 microns.
Still another approach to providing nanowires is to utilize an
anodic alumina substrate to prepare carbon nanotubes inside the
pores of the anodic alumina by the carbonization of propylene
vapor. One problem with such a gas phase reaction approach is that
it is relatively difficult to generate dense continuous
nanowires.
To overcome the problems of filling pores in a template, high
pressure-high temperature material injection techniques have been
used. In these techniques, a molten metal is injected into
relatively small diameter pores of a template to make nanostructure
composites. In one technique, a hydrostatic press provides a
relatively high pressure to inject metals such as indium (In),
gallium(Ga) or mercury (Hg) into the pores of the substrate. This
technique may also be used to fill a single glass nanotube having a
diameter of about 100 nm with a molten metal such as bismuth (Bi).
The technique may also be used to fill porous anodic alumina with
channel diameters larger than 200 nm with various metal melts. One
equation which may be used to compute the rate at which a molten
metal can be injected into a pore of a template is shown in
Equation 1: ##EQU1##
in which:
l(t) is an injection length at time t;
P.sub.a is an external pressure;
.gamma..sub.lv is a liquid-vapor surface tension;
.theta. is a contact angle between the liquid and a wall of the
pore;
r is a pore radius;
.eta. is a viscosity of the liquid; and
.epsilon. is a coefficient of slip of the liquid.
The contact angle .theta. may be computed using Young's equation
which may be expressed as:
in which .gamma..sub.sv and .gamma..sub.sl are the solid-vapor and
solid-liquid surface tensions, respectively. For a ceramic/metal
melt system, the difference between the solid-vapor surface tension
.gamma..sub.sv and the sold-vacuum surface tension .gamma..sub.so
is negligible. Through simple thermodynamic calculations, the
following relation for the contact angle .theta. is reached:
in which V.sub.sl and V.sub.ll are the solid-liquid and
liquid-liquid interaction energies, respectively.
In metallic liquids the liquid-liquid interaction energy V.sub.ll
is relatively strong. Thus, when injecting metallic liquids in
prior art techniques the contact angle .theta. was assumed to be
close to 180.degree..
With such an assumption, the external pressure needed to drive the
molten metal into a channel with diameter D is
P.sub.a.gtoreq.-4.gamma..sub.lv /D.
As an example, the solid-liquid surface tension .gamma..sub.sl of
liquid bismuth is about 375 dyn/cm. Assuming the contact angle
.theta. is 180.degree., a pressure of 1,150 bar is needed to fill a
pore having a diameter of about 13 nm. Such a high pressure can be
achieved by a hydrostatic press. The melting temperature of
bismuth, however, is 271.5.degree. C. Thus, to fill a pore having a
diameter of about 13 nm, a hydrostatic press must provide a
pressure of 1,150 bar at a temperature of at least 271.5.degree. C.
Due to the combination of high pressure and high temperature, and
the corresponding problems associated with operating hydrostatic
equipment at such high temperatures and pressures, it was
heretofore not practically possible to inject liquids and, in
particular, liquid metals into pores having relatively small
diameters.
Moreover, even if relatively small diameter pores in anodic alumina
could be filled, as explained above, the anodic alumina itself is
typically unsuitable as a template for quantum wires, due to the
lack of a systematic technique for controlling the diameter, a
length and packing density of the pores in the anodic alumina.
It would, therefore, be desirable to provide a technique for
fabricating an array of nanowires having a relatively small
diameter, a relatively close spacing and a relatively long length.
It would also be desirable to provide a technique for fabricating
nanowires which does not depend upon the selection of chemical
precursors. It would also be desirable to provide a technique which
can be used to fabricate continuous wires having a relatively long
length and which does not require high pressure injection of molten
materials at relatively high temperatures. It would also be
desirable to provide a template having pores therein with pore
diameters which are relatively uniform. It would also be desirable
to provide a technique for filling substrate pores having
relatively small diameters. It would also be desirable to provide a
technique for systematically controlling the pore diameter, pore
length and center-to-center spacing of pores in an anodic aluminum
oxide template.
SUMMARY OF THE INVENTION
In accordance with the present invention, an array of nanowires
includes a substrate having a plurality of non-interconnected pores
each of the pores having pore diameter which does not vary by more
than one hundred percent and a material continuously filled in each
of the plurality of pores of the substrate wherein the material has
characteristic such that the material can become a quantum wire
having a length not less than three microns. With this particular
arrangement, an array of non-interconnected nanowires which can be
used in a semiconductor device, an optical device or a
thermoelectric device is provided. The substrate may, for example,
be provided from a metal such as aluminum or an aluminum alloy in
sheet or metal form having a surface layer of aluminum oxide
thereon. Alternatively, the substrate may be provided from a
mesoporous material such as a material from the
silicate/aluminosilicate mesoporous molecular sieves. The material
disposed in the pores may be provided as bismuth (Bi), or any other
material capable of becoming a quantum wire. The substrate pores
are provided having a wall composition or a surface property which
reduces the contact angle between the material continuously filling
the pores and the pore wall. Thus, the substrate pores can be
filled utilizing relatively little, if any, pressure.
In accordance with a further aspect of the present invention, a
method for providing a substrate having pores with walls having
reduced contact angles includes the step of treating the pore wall
with an acid solution to change at least one of a pore wall
composition and a pore wall surface property. With this particular
arrangement, a substrate having pores which can be filled without
high pressure injection of a molten material at a relatively high
temperature is provided. In one embodiment, the substrate is
provided as an anodic aluminum oxide film, which is prepared by the
anodic oxidation of aluminum in an acidic electrolyte. The
electrolyte solution is selected to provide an anodic aluminum
oxide film having a pore with a wall surface composition which is
different than aluminum oxide. Thus, the pore wall composition or
properties are modified during an anodization process. The modified
pore walls result in a contact angle between the pore wall and a
filling material which allows the pore to be continuously filled
with a material without the use of high pressure injection
techniques. Alternatively, the pore wall composition or surface
properties may be modified after the anodizing process by applying
a solution of H.sub.2 SO.sub.4 to the pore wall to thus change the
composition or surface properties of the pore walls to provide the
pore walls having a contact angle which allows molten material to
fill the pores without the use of high pressure injection
techniques. Alternatively still, the composition or surface
properties of the pore walls may be modified by depositing a
desired surface species on the pore wall by a vapor deposition
technique, for example. In another embodiment, the substrate is
provide as mesoporous MCM-41. The pores in the mesoporous material
may also be treated such that the contact angle between the pore
walls and the material filling the pores allows the material to be
drawn into the pores with little or no pressure.
In accordance with a still further aspect of the present invention
a technique for fabricating nanowires includes the steps of
treating the pores of an anodic aluminum oxide film to improve a
contact angle of a pore wall, melting metal under vacuum and
injecting the molten metal under pressure into the pores of the
anodic aluminum oxide film to produce continuous nanowires. With
this particular arrangement, a dense array of continuous nanowires
having relatively small diameters which can be utilized in
transport-related applications is provided. In one embodiment, the
anodic aluminum oxide film has a plurality of pores formed therein
and the technique is used to provide a dense array of nanowires.
Thus, the process of the present invention can be utilized to
generate large areas of highly regular and densely-packed nanowire
arrays. Moreover, the process does not require clean room
technology as is necessary for fabrication of quantum well
superlattice systems. Therefore, a relatively simple and
inexpensive technique for fabrication of densely-packed arrays of
continuous nanowires is provided. Another advantage of this
technique is that it can be applied to a wide range of materials
including low melting temperature metals, alloys, semiconductors,
and organic polymer gels and thus the technique is versatile. In
one particular application an array of bismuth nanowires having
average wire diameters as small as 13 nm, lengths of 30-50 .mu.m,
and a packing density greater than 4.6.times.10.sup.10 cm.sup.-2 is
provided.
In accordance with still a further aspect of the present invention
a method to systematically change the channel diameter and channel
packing density of anodic aluminum oxide film includes the steps of
anodizing an aluminum substrate with a particular one of a
plurality of electrolytes at a predetermined voltage level, a
predetermined temperature and a predetermined current and exposing
pores in the anodized aluminum substrate to an acid which modifies
either the composition or a surface property of a surface of the
pore walls. With this particular arrangement, a systematic method
for providing aluminum oxide film having particular characteristics
is provided. The method can be used to provide, for example, an
anodic aluminum oxide film having a particular pore diameter in the
range of pore diameters extending from about 8 nm to about 200 nm.
Moreover, with this technique, the pore diameters do not vary by
more than one-hundred percent. Thus, with this technique, an anodic
aluminum oxide film having an average pore diameter of 8 nm and
having a desired channel length and structural regularity can be
provided.
The solid-liquid surface tension .gamma..sub.sl, depends on the
surface properties of the solid in which the pores are formed and
is not always small compared to the liquid-liquid surface tension
.gamma..sub.ll. By changing the composition or a surface property
of the pore wall, a desired wall surface for individual liquids can
be produced, thereby reducing the contact angle .theta.. It has
been recognized that in the pressure injection process, the contact
angle .theta. plays an important role. Specifically, if the contact
angle .theta. is less than 90.degree., the capillary pressure
4.gamma..sub.lv cos .theta./D is positive, so that this pressure
itself is able to drive the liquid into the pores even when the
pores are provided having a relatively small diameter. If the
contact angle .theta. is greater than 90.degree., however, an
external pressure higher than -4.gamma..sub.lv cos .theta./D is
needed. For an anodic alumina template the portion of alumina that
is closed to the internal surface of the channels is contaminated
by anions from the anodizing electrolyte. This means that
solid-liquid surface tension .gamma..sub.sl depends on both the
specific metal melt and the electrolyte type. Thus by appropriately
selecting a suitable electrolyte the contact angle can be reduced
and molten materials may be driven into pores having diameters at
least as low as 8 nm with relatively low or no pressure. In one
experiment, an anodic alumina template was prepared using a
sulfuric acid solution and molten bismuth was successfully driven
into pores with diameters as small as 13 nm at a pressure of 315
bar. The sulfuric acid electrolyte used for bismuth may also be
appropriate for other molten metals. It is recognized, however,
that the use of an anodizing electrolyte or a pore etching solution
to adjust the pore wall composition or surface property may not be
applicable to every molten material or even to every molten metal.
If a suitable acid solution cannot be found for injecting a
particular molten material of interest, a vapor deposition of a
desired surface species before the pressure injection process may
be performed in place of or in addition to, the use of an anodizing
electrolyte or a pore etching solution to thus control the
composition or surface properties of the pore wall to provide a
pore wall surface which favors the interaction between the pore
wall surface and molten material of interest. This technique allows
the use of a modest pressure to drive a molten or liquid material
into the pores and thus existing manufacturing and processing
equipment can be used.
In accordance with a still further aspect of the present invention,
a method to control the channel diameter of an anodic alumina film
and the ratio of a pore diameter to the cell size during and after
an anodizing process is described. With this particular technique,
a method for controlling the pore structure of an anodic alumina
film is produced. A vacuum melting and pressure injection process
can be used to then fill an array of densely packed pores to thus
generate continuous and dense nanowires useful in many electronic
applications. Due to the high thermal and chemical stability of the
anodic aluminum oxide film, the pressure injection process can be
applied to other low melting temperature metals, semiconductors,
alloys, and polymer gels. In one particular experiment an array of
bismuth nanowires having an average diameter of about 13 nm, a
length of about 30 .mu.m and a 7.1.times.10.sup.10 cm-.sup.-2
packing density was fabricated. Moreover, the individual wires were
provided having a single crystal lattice structure.
The nanowire array composites fabricated in accordance with the
techniques of the present invention find applicability in a wide
range of fields including but not limited to use in electronic
devices, photonics, high Tc superconductivity, thermoelectricity,
chemical gas sensors and chemical gas separation. In particular,
the 1D quantum wire systems find application in a wide range of
technical fields of use.
BRIEF DESCRIPTION OF THE DRAWINGS
The foregoing features of this invention as well as the invention
itself may be more fully understood from the following detailed
description of the drawings in which:
FIG. 1 is a highly diagrammatical view of a substrate having a
plurality of pores therein;
FIG. 1A is a cross-sectional view taken through lines 1A--1A in
FIG. 1;
FIGS. 2-2B are a series of diagrams illustrating the steps in the
fabrication of a metal nanowire array composite;
FIG. 3 is a block diagram of a system for anodizing and
electrochemical polishing a substrate;
FIG. 4 is a schematic diagram of a system for the vacuum melting
and pressure injection of molten metal into the pores of a
substrate;
FIGS. 5-5B are a series of images generated with a scanning
electron microscope (SEM) of the topside surface of an anodic
aluminum oxide film prepared by a 20 wt % H.sub.2 SO.sub.4 solution
at 20V (FIGS. 5 and 5A) and 15V (FIG. 5B);
FIG. 6 is an SEM image of a backside surface of an anodic aluminum
oxide film taken before removal of a barrier layer;
FIG. 6A is an SEM image of a backside surface of an anodic aluminum
oxide film after removal of a barrier layer;
FIGS. 7-7B are a series of images taken with a transmission
electron microscope (TEM) of a cross-section of an anodic aluminum
oxide film after ion milling;
FIG. 7C is a plot of the number of pores versus average pore
diameter for the anodic aluminum oxide film of FIG. 7;
FIG. 7D is a plot of the number of pores versus average pore
diameter for the anodic aluminum oxide film of FIG. 7A;
FIG. 7E is a plot of the number of pores versus average pore
diameter for the anodic aluminum oxide film of FIG. 7B;
FIGS. 8-8A are a series of SEM images of a top surface of an anodic
aluminum oxide film having pores filled with bismuth;
FIGS. 9-9A are a series of SEM images of a bottom surface of an
anodic aluminum oxide film having bismuth filled pores after the
removal of a barrier layer;
FIGS. 10-10B are a series of SEM images of metal nanowires after
the anodic aluminum oxide film was partially dissolved from a
bottom surface of the film;
FIG. 11 is a schematic diagram of a system for fabricating
nanowires;
FIG. 12 is a plot of an x-ray diffraction (XRD) pattern of a
bismuth nanowire array made from an anodic aluminum oxide film;
FIG. 13 is a plot of an XRD pattern for an array of bismuth
nanowires formed in an anodic aluminum oxide film;
FIG. 14 is a plot of an XRD pattern for a bismuth nanowire array
formed in an anodic aluminum oxide film;
FIG. 15 is a plot of optical transmission spectra of an anodic
aluminum oxide film and a plurality of bismuth nanowire arrays
formed in an anodic aluminum oxide film;
FIG. 16 is a plot of optical transmission spectra for an anodic
aluminum oxide film and an array of bismuth nanowires formed in an
anodic aluminum oxide film.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
Terminology
Before describing the processing and structures utilized to
fabricate an array of nanowires, some introductory concepts in
terminology are explained. As used herein, the term nanometer size
range is used to describe a size range from about one nanometer to
about 500 namometers.
Quantum wires can be generally defined as including regions in
which the charge carriers are quantum confined in two dimensions
orthogonal to the axis of the wire. Quantum wire devices consist of
a collection of particles each having a resonance cavity so small
that quantum confinement effects are very pronounced. For the
device to be effective, therefore, there must be a very high degree
of size uniformity of the particles making up the quantum dot or
quantum wire device so that each has substantially identical
electronic and optical properties.
Quantum wire devices provide the building blocks for digital
nanoelectronic circuits which refer to an integrated circuit
technology which permits down scaling to be carried beyond to what
is currently achievable. Quantum wire devices also provide the
basic structure for thermoelectric devices.
Referring now to FIG. 1, an array of nanowires 10 includes a
substrate 12 having a length L.sub.1, a width W.sub.1 and a
thickness T.sub.1. Substrate 12 is provided having a plurality of
non-interconnected pores 14a-14n generally denoted 14. The pores 14
are non-interconnected in the sense that the pores 14 do not
contain apertures or openings to any other of the pores 14. For
example pore 14a does not contain any opening or aperture through
which material could move into pore 14b. Each of the pores are
continuously filled with a material which provides a corresponding
nanowire 15a-15n generally denoted 15.
In this particular embodiment, each of the wires 15 are provided
having a diameter in the range of about 80 angstroms (.ANG.) to
about 300 nanometers (nm) with the diameter of any single wire and
thus any single pore 14 not varying by more than one hundred
percent along the length of the wire. In a preferred embodiment,
substrate 12 may be provided from any one of a plurality of
materials having a plurality of pores therein. For example,
substrate 12 may be provided as an anodic aluminum oxide film. As
is known, porous anodic alumina has an array of hexagonally packed
nanometer-sized pores. As will be described in detail herein below,
certain processing steps can be performed to tailor the diameter,
length and packing density of the pores in the anodic aluminum
oxide film. The substrate is preferably provided having a thickness
in the range of about 200 nm to about fifty micrometers (.mu.m).
Anodic aluminum oxide film is desirable because of its geometry and
also because its wide band gap energy and low thermal conductivity
(<4 watt.multidot.cm.sup.-1 K.sup.-1) are desirable when the
anodic aluminum oxide film acts as a template for quantum wire
systems of interest for electronic or optical applications.
Alternatively, substrate 12 may be provided from a mesoporous
material such as mesoporous silica also having a thickness in the
range of about 200 nm to about fifty microns. Mesoporous material
may be provided, for example, as MCM-41. Alternatively still
substrate 12 may be formed from an array of glass nanotubes or from
an organic matrix.
Regardless of the particular substrate material, nanowires 15 are
provided by continuously filling the pores 14 in the substrate 12
with a preferred material. The particular material with which pores
14 are filled may be selected in accordance with a variety of
factors including but not limited to the particular application for
which the nanowire array will be used. For example, when the array
of wires 15 are to be used in a thermo-electric device, then it is
desirable to provide the wires from a material such as bismuth
(Bi). Alternatively, when it is desirable to use an array of
nanowires 14 in a superconductivity, electronic, optical or
chemical gas application it may be preferable to form the nanowire
array from materials other than Bi. Those of ordinary skill in the
art will appreciate and recognize how to select a particular
material for a particular application.
After providing substrate 12 having a plurality of
non-interconnected pores 14 with average diameters less than nine
nanometers and diameter variations of less than one hundred
percent, the pores may be filled with a material using one of a
variety of filling techniques. One particular technique for
providing a substrate having an array of regularly spaced pores
which may also have relatively small pore diameters will be
described herein below. With respect to filling techniques the
pores may be filled with a material using an evacuation and
pressure filling technique to be described below or alternatively
the pores may be filled using an electro deposition technique or a
vapor deposition technique including a chemical vapor deposition
technique. Prior to filling the pores, however, the pore walls are
treated to such that the contact angle between the pore wall and
the filling material is reduced.
Referring now to FIG. 2, a porous substrate 20 includes an anodic
aluminum oxide film 22 disposed over a barrier layer 24 on top of
an aluminum carrier 26. Anodic aluminum oxide film 22 has a
plurality of pores 28 provided therein. Pores 28 are a result of
the anodizing process and in this particular embodiment, have a
diameter typically in the range of about eight to about one hundred
nanometers with a diameter of variation along a length of the pore
of no more than one hundred percent and preferably no more than
fifty percent and even more preferably no more than twenty percent.
Pores 28 are separated by barrier walls 30 and are separated from
substrate 26 by a barrier layer 24.
Referring now to FIG. 2A, a pore filling material which may, for
example, be provided as a metal such as bismuth (Bi) 32
continuously fills a predetermined region in each of the pores 28.
A portion 33 of the filling material 32 merges with barrier layer
24. The layer 33 and carrier 26 are then removed in a manner to be
described below to provide an array of nanowires 34 as shown in
FIG. 2B. Nanowire array 34 is provided from a plurality of walls 36
which define openings in the substrate. The openings are filled
with the wire material 38. Thus an array of equally spaced wires 38
each having a substantially constant diameter is provided.
EXAMPLE
An aluminum substrate was used in the fabrication of an array of
bismuth nanowires using a preferred process of the present
invention. First, an aluminum substrate was prepared in the
following manner.
In one particular embodiment, to provide the an anodic aluminum
oxide film such as film 22 described above in conjunction with FIG.
2, an aluminum (Al) foil having a purity of about 99.997% was
divided into a plurality of pieces of predetermined shape. The Al
foil may be provide as commercially available Al foil of the type,
provided for example, from Alfa AESAR and cut into pieces having a
rectangular shape with a length typically of about 25 millimeter
(mm) and a width typically of about 10 mm. Those of ordinary skill
in the art will appreciate of course that the Al foil may be
provided having a shape other than rectangular and may be provided
any length, width or diameter appropriately selected for a
particular application and that it is not a necessary to the
invention that the Al foil be provided in sheet form.
In the case where Al foil is used, the Al foil are flattened. This
may be accomplished, for example, by pressing the foil to a flat
aluminum substrate using a cold press apparatus. The Al foil is
pressed to the substrate, operating at a pressure of about
1.7.times.10.sup.3 bar.
The next step is to polish the foil. The foil may be polished using
a combination of mechanical and electrochemical polishing
techniques. Mechanical polishing may be accomplished using, for
example, a three micron diamond paste and then using Mastermate.
The mechanically polished foil normally has a shiny finish, but it
is not perfect over the whole surface. After mechanical polishing,
the foil is washed in acetone and sonicated for a time typically of
about thirty minutes. The foil is then air dried. It should be
noted that if the foil is initially provided as a flat foil having
a relatively smooth substrate, the above-described flattening and
mechanical polishing steps may be omitted and processing of the
foil begins by electrochemical polishing of the Al foil.
Prior to the electrochemical polishing step, the substrate is
calcined in air at a temperature typically of about 350.degree. C.
for a time of approximately thirty minutes to provide the substrate
having a substantially uniform surface oxide layer. Next the foil
is electrochemically polished. The step of electrochemical
polishing includes the steps of placing the aluminum substrate in
an electrolyte provided from an acid solution, providing the
aluminum substrate as an anode, providing a cathode and applying a
voltage in a predetermined voltage range.
Referring briefly to FIG. 3, an apparatus for polishing the foil
includes a container 42 having a mixed solution of acids as the
electrolyte disposed therein. The electrolyte solution is
appropriately selected for electrochemical polishing of the foil.
For example, the composition of the acid solution may be provided
as: 95 vol % phosphoric acid solution (85 wt %)+5 vol % sulfuric
acid solution (97 wt %)+20 g/l chromium oxide (CrO.sub.3).
The temperature of the electrolyte is maintained at a temperature
of about 85.degree. C. The Al film serves as an anode 46 and a Pt
foil or a graphite electrode may be used to provide a cathode 48. A
power supply 50 provides a substantially constant voltage in the
range of about 20 V to about 24 V during the polishing. The
polishing time is controlled to be between about one minute to
about three minutes depending upon the original surface roughness
of the foil 46.
After polishing, the foil 46 is removed from the electrolyte 44,
washed with de-ionized water, and then dried in air. The substrate
46 is next de-smudged in a chromic acid solution (CrO.sub.3 : 45
g/l, H.sub.3 PO.sub.4 3.5 vol %) at a temperature of about
95.degree. C. for several minutes to dissolve the surface oxide on
the substrate 46 immediately before the anodizing process. The
substrate 46 is then washed again with de-ionized water and dried
in flowing air. The surface of the Al substrate 46 normally has a
relatively smooth finish after the electrochemical process.
It should be recognized that other electrochemical techniques, well
known to those of ordinary skill in the art may also be used to
provide the Al film having an appropriately smooth surface
finish.
After preparing the Al film as discussed above, the film is subject
to an anodizing process. In the anodizing process, a strong acid
solution is preferably used as the electrolyte. Depending upon
desired channel diameters and wall compositions of the anodic
alumina films, one of three types of electrolytes (H.sub.2 SO.sub.4
solution, oxalic acid (H.sub.2 C.sub.2 O.sub.4) solution, and
H.sub.3 PO.sub.4 solution) may be used.
When a solution of H.sub.2 SO.sub.4 is used as the electrolyte, the
solution is provided having a weight of H.sub.2 SO.sub.4 in the
range of about 5% to about 40% and preferably a weight of H.sub.2
SO.sub.4 in the range of about 15% to about 20%. Typically, when it
is desirable to provide an anodic alumina film having pore
diameters 30 nm or less and relatively small cell sizes are desired
or required, a solution having a weight of H.sub.2 SO.sub.4 in the
range of about 15% to about 20% is preferably used as the
electrolyte.
When a solution of H.sub.2 C.sub.2 O.sub.4 is used as the
electrolyte, the solution is provided having a weight of H.sub.2
C.sub.2 O.sub.4 in the range of about 1% to about 8% and preferably
a weight of H.sub.2 SO.sub.4 in the range of about 3% to about 5%.
Typically, when it is desirable to provide an anodic aluminum oxide
film having pore diameters in the range of about 30 nm to about 80
nm, a solution of having a weight of H.sub.2 C.sub.2 O.sub.4 of
about 4% is preferably used as the electrolyte.
When a solution of H.sub.3 PO.sub.4 is used as the electrolyte, the
solution is provided having a weight of H.sub.3 PO.sub.4 in the
range of about 1% to about 40% and preferably a weight of H.sub.2
SO.sub.4 in the range of about 4% to about 8%. Typically, when it
is desirable to provide an anodic aluminum oxide film having pore
diameters in the range of about 80 nm to about 200 nm, a solution
of 4-8 wt % H.sub.3 PO.sub.4 is preferably used as the
electrolyte.
The portion of anodic aluminum oxide film surrounding the pores
(i.e. the pore wall surfaces) are contaminated by anions from the
electrolyte and it has been recognized that changing the
electrolyte type can change the composition of the pore wall. Thus,
the composition of the anodic aluminum oxide film, and in
particular the pore wall surfaces of the anodic aluminum oxide film
may be important. Thus, the H.sub.2 C.sub.2 O.sub.4 solution can be
used to prepare anodic alumina films having pores with diameters
smaller than 20 nm, and the H.sub.3 PO.sub.4 solution can be used
to prepare anodic alumina films having pores with diameters as
small as 30 nm.
Next, the Al film is anodized under constant voltage conditions. It
has been found that using constant voltage conditions in the
anodizing process results in the repeated generation of anodic
aluminum oxide films having a relatively regular array of pores.
The voltage determines the cell size of the film. All three
electrolytes were found to obey the following empirically derived
equation:
in which:
C is the cell size in nanometers;
V is the anodizing voltage in volts; and
m is a constant in the range of 2.0 to 2.5.
For a solution of H.sub.2 SO.sub.4, a voltage in the range of about
5 volts to about 30 volts may be used. If the solution is provide
having a weight of H.sub.2 SO.sub.4 in the range of about 15% to
about 20% a voltage in the range of about 13 volts to about 25
volts is preferably used. If the voltage is selected to be higher
than 30 volts, a relatively high current flow results which further
results in generation of an anodic aluminum oxide film which does
not have a suitably uniform structure. If the voltage is lower than
about 5 volts, a relatively low current flow results which also
results in generation of an anodic aluminum oxide film which does
not have a suitably uniform structure.
For a solution of H.sub.2 C.sub.2 O.sub.4 a voltage in the range of
about 5 volts to about 120 volts may be used. For a solution of
having a weight of H.sub.2 C.sub.2 O.sub.4 of about 4%, a voltage
between 30 to 60 is preferred.
For a solution of H.sub.3 PO.sub.4, a voltage in the range of about
5 volts to about 200 volts may be used. For a solution having a
weight of H.sub.3 PO.sub.4 in the range of about 4% to about 8%, a
voltage in the range of about 15 volts to about 120 volts can be
applied depending on the actual acid concentration. Typically,
however, to generate a reasonable anodizing current, a voltage
higher than 25 volts may be needed. It was found that higher
anodizing voltages within the operating range typically generate
better pore structures for all three electrolytes that were
employed.
In the anodizing process, the current plays an important role in
generating films with a regular structure. Normally a current in
the range of 1 to 200 mA/cm.sup.2 and preferably between 1 to 80
mA/cm.sup.2 and even more preferably between 5 to 40 mA/cm.sup.2 of
substrate surface is desired. If the current is too large, the film
will grow very fast and the structure that is generated is not
uniform. If the current is too small, the film growth will be very
slow and a very long anodizing time is necessary to generate a film
with the desired thickness. The anodizing current depends on the
selected voltage, electrolyte type and electrolyte temperature.
The electrolyte concentration is another important parameter in
controlling the film growth. Relatively high electrolyte
concentrations increase the anodizing current for the same
anodizing voltage. Higher concentrations generally favor slightly
larger pore diameters even though this effect was not very
significant for all three of the electrolytes.
The electrolyte temperature is also an important parameter in
controlling film growth. First, the activity of the acid strongly
depends on the temperature. For an H.sub.2 SO.sub.4 solution, the
pore diameter an anodic aluminum oxide film is dependent on the
electrolyte temperature. For very low temperature, e.g. 0.degree.
C., a very small pore diameter will be generated. The pore diameter
increases with increasing temperature. A suitable electrolyte
temperature range is 0-10.degree. C. A similar behavior was
observed for the H.sub.2 C.sub.2 O.sub.4 solution. Since the pore
diameter generated with H.sub.2 C.sub.2 O.sub.4 is normally larger
than that with H.sub.2 SO.sub.4, the influence of the electrolyte
concentration was not that important. Second, the electrolyte
temperature affects the anodizing current. For all three
electrolytes, the current increases slightly when the temperature
increases. So the temperature of the electrolyte is a useful
processing parameter to control both the film growth rate and the
ratio of the channel diameter to the cell size.
Many anodic aluminum oxide films prepared as described above did
not have a smooth channel wall surface. It is desirable, however to
provide pores having relatively smooth pore wall surfaces since
this may facilitate filling the pores with a liquid using a filling
technique such as the injection technique to be described below in
conduction with the step of filling the pores with a liquid metal.
The channel diameter can be adjusted by etching with an acid
solution. The final channel diameter of the film depends on both
the composition of the etching solution and the etching time. The
4-5 wt % H.sub.3 PO.sub.4 solution is a widely used etching
solution and is also an efficient etching solution. Another etching
solution which used in this experiment was a solution of 20 wt %
H.sub.2 SO.sub.4, which was found to work much slower than the 4-5
wt % H.sub.3 PO.sub.4 solution.
In this experiment, it was discovered that the surface properties
of the channel wall could be strongly affected by the acid etching.
When pore enlargement was performed using the H.sub.3 PO.sub.4
solution for films prepared with a solution of 20 wt % H.sub.2
SO.sub.4 as the electrolyte, it was very difficult to drive molten
Bi into the channels at a pressure of 315 bar. When pore
enlargement of the same film was performed using a H.sub.2 SO.sub.4
solution, however, molten Bi was driven into the small channels at
the same pressure of 315 bar even though the former film actually
had a slightly larger average channel diameter.
It is believed that the anions from the etching solution changed
the composition of the channel walls, which was a gel-like material
in the as-prepared anodic alumina films. For films prepared by the
H.sub.2 C.sub.2 O.sub.4 solution, the pore enlargement process was
performed according to the following sequence: first the film was
dipped into the H.sub.3 PO.sub.4 solution to generate the desired
channel diameter, then it was washed in de-ionized water and dipped
into the H.sub.2 SO.sub.4 solution to change the surface properties
of the channel.
The pores were filled with a wire material using a vacuum melting
and pressure injection processes next described. In order to
stabilize the film structure, the anodic alumina films were first
calcined at 350.degree. C. for 1 hour in air. The small channels of
the host film were then filled with Bi metal by the vacuum melting
and high pressure injection processes. The low melting temperature
of Bi (T.sub.m =271.5.degree. C.) and the high thermal stability
(up to 800.degree. C.) and high rigidity of the anodic alumina film
make these processes possible.
As may be seen with reference to FIG. 4, the porous film, which was
kept on the aluminum substrate, was placed inside a high pressure
reactor chamber and surrounded by high-purity Bi pieces. The
reactor chamber was first evacuated to pressure of about 10.sup.-2
mbar and heated to a temperature slightly lower than the melting
temperature T.sub.m of Bi to degas the porous film.
The chamber was then heated to a predetermined temperature at a
ramp rate of 2.degree. C./min heating rate. In this particular
example, the predetermined temperature was selected to be about
250.degree. C. A temperature soak at 250.degree. C. for 3 hours
were found to be sufficient to accomplish the goal of degassing the
porous film.
After the film was fully degassed, the temperature of the reactor
was increased above the T.sub.m of Bi, and a temperature of
325.degree. C. was used in the experiments. The higher temperature
above T.sub.m helped to slightly reduce both the surface tension
and the viscosity of the molten metal. High pressure argon gas was
introduced into the reactor chamber to drive the molten Bi into the
evacuated channels. The injection time was found to depend on the
pressure, channel length, surface properties of the channel wall,
and the individual liquid. In these experiments, 3 to 8 hours were
used for the injection time.
After the injection was completed, the reactor was slowly cooled
down to room temperature, normally at a cooling rate of
0.5-1.degree. C./min in these experiments. The impregnated Bi was
allowed to solidify and to crystallize inside the nanochannels. The
slow cooling was found to be important for generating single
crystal Bi nanowires. Finally the pressure was slowly released.
Referring briefly to FIG. 4, apparatus 52 effective for performing
the above described vacuum melting and pressure injection of a
material, such as Bi for example, into pores of a substrate such as
an anodic aluminum oxide film, includes a chamber 54. Chamber 54
may be provided as a high pressure chamber (which also serves as a
vacuum chamber) having a heater 56 coupled thereto. A sample 58 is
disposed in chamber 54 and is surrounded by material 60 to be
injected into a porous surface of the sample 58.
Means 62 for pressuring chamber 54 is here provided by from a gas
source which in one particular embodiment may be provided as source
of pressurized gas such as argon (Ar) gas and in particular an
argon gas tank capable of providing a pressure of about 6000 pounds
per square inch. Also coupled to chamber 54 is a means 64 for
generating a vacuum within chamber 54. In one particular
embodiment, means 64 includes a vacuum pump for evacuating the
chamber 54 to a level of about 10.sup.-2 torr. Apparatus 52 may be
provided for example as a commercially available high pressure
reactor. In general overview, apparatus 52 is operated in a manner
similar to that described above to inject a desired material into
the pores of the sample 58.
As for the wetting process, in one experiment in which molten Bi
was injected into the anodic aluminum oxide film prepared by a
H.sub.2 SO.sub.4 solution at a pressure of about only 10 bar. It is
possible that the contact angle between the liquid Bi and the
channel walls was such that the capillary pressure itself was able
to drive the liquid Bi into the small nanochannels. In another
experiment, pore enlargement was performed using 4 wt % H.sub.3
PO.sub.4 solution for the anodic alumina prepared by an H.sub.2
SO.sub.4 solution, and the average channel diameter of the film
after pore enlargement was 35 nm. However, the molten Bi melt was
not be able to be injected into the 35 nm channels of the H.sub.3
PO.sub.4 solution etched film at the same pressure (315 bar) at
which we filled the 13 nm channels. This indicated the importance
of the wetting processes used to modify the composition or a
surface property of a pore wall to provide a contact angle which
allows injection of a liquid metal at relatively low injection
pressure.
After the liquid Bi was injected into the pores of the anodic
aluminum oxide film, the processed film (sometimes referred to
hereinbelow as a "sample") was mechanically extracted from the
metal piece. Since the porous film remained on the Al carrier
during the processing, it was not difficult to separate the sample
from the Bi metal. Normally there was a crack between the film and
the metal piece. This crack is believed to be due to the
incompatibility of the thermal expansion rate between the anodic
alumina and the Bi during the cooling process.
When a shear force was applied between the sample and the Bi metal,
the sample separated from the metal piece as a whole. Most of the
time, there were no Bi metal pieces attached to the film surface
after separation. After the sample was extracted from the metal
piece, the Al substrate was etched away by an amalgamation process.
Here, a 0.2 M mercuric chloride (HgCl.sub.2) solution was used. It
was observed that a thin layer of Bi (e.g. layer 33 in FIG. 2A)
existed between the film and the Al carrier after the Al carrier
was removed.
The Bi layer 33 (FIG. 2A) was polished away mechanically using 50
nm .gamma.-Al.sub.2 O.sub.3 particles, which were found to be a
good polishing agent for this process, since 50 nm .gamma.-Al.sub.2
O.sub.3 particles has almost the same hardness as the anodic
alumina film and did not damage the film structure. After the
polishing process, the film was put on a glass slide using molten
paraffin with the upper surface of the anodic alumina film facing
the glass slide so that it could be protected by the solidified
paraffin. The glass slide was then dipped into a 4 wt % H.sub.3
PO.sub.4 solution to dissolve the barrier layer. The etching time
depended on the thickness of the barrier layer h which is
proportional to the anodizing voltage V, through the empirical
relation: h(nm).apprxeq.V(volts). The etching rate was about 1
nm/minute. It was found that there was a short distance between the
end of the majority of Bi wires and the backside surface of the
original film (see FIG. 2A), and this distance is attributed to the
shrinkage of the Bi during the liquid to solid phase transition.
This empty portion of the channels can be easily removed by
chemical etching or mechanical polishing. A schematic of the final
composite structure is shown in FIG. 2B.
The anodic alumina film was evaluated using a high resolution
scanning electron microscope (SEM JOEL 6320) and a transmission
electron microscope (TEM, JOEL 20OCX). Ion milling (6 keV argon
ion) was used to thin the film to a thickness less than 100 nm so
that a TEM picture of the film could be taken. The Bi filled anodic
alumina firm was characterized by X-ray diffraction (XRD) using a
Siemens D5000 diffractometer (45 kV-40 mA, Cu--K.alpha.). The Bi
nanowires were evaluated by SEM. The band gap energy of the Bi
nanowires was determined by taking optical transmission spectra of
the composite films using an UV-Visible-IR spectrophotometer.
FIGS. 5-10 are a series of images produced using SEM and TEM to
show the structure of the substrates and nanowire arrays fabricated
in accordance with the techniques of the present invention.
Referring now to FIGS. 5-5B, SEM images of the surface of the
anodic alumina film are shown. SEM provides a very efficient tool
for evaluating the surface features of the anodic alumina film.
Since the film is insulating, the film was coated with a thin layer
of Au/Pd before the SEM experiment, which helped to get high
resolution images. It was found that the average diameter of the
channel entrances was slightly larger than the expected value
(there were empirical relations between the average channel
diameters and the anodizing voltages in the literature). This is
consistent with previous results in the literature. SEM was also
used to evaluate the backside of the film after the barrier layer
was removed. The regularity of the pore structure on the backside
of the film was found to be better than the topside surface, Some
of the SEM images of the film backside are presented in FIGS. 6 and
6A. The film thickness was also determined from SEM images.
In order to get more accurate data on the channel size
distributions of the anodic alumina films, the films were thinned
from both sides by Ar ion milling (argon ions at 6 KeV). The pore
structure of the film was then evaluated by TEM. This gave us an
image of the cross section of the film, and a more faithful pore
size distribution than SEM images because the channel entrances
normally have different diameters from the majority of channel
lengths inside the film.
Shown in FIGS. 7-7B are TEM images of anodic alumina films with
average channel diameters of 56, 23, and 13 nm and channel
densities of 7.4.times.10.sup.9, 4.6.times.10.sup.10 and
7.1.times.10.sup.10 cm.sup.2, respectively.
FIGS. 7C-7E, show the distributions of the average channel diameter
of these films, we can see that the anodic alumina films which we
made have a highly regular structure and a narrow channel size
distribution. The channel diameters determined in this method were
normally smaller than the diameters of the channel entrances which
were determined by SEM images of the film surface.
The crystalline structure of the Bi nanowire array composites were
studied by X-ray diffraction. Before the XRD experiment, the Bi
filled films were first extracted from the Al substrate. Then the
alumina barrier layer on the backside of the film was removed,
which helped to assure that there were no contributions to the
diffraction pattern from Bi particles stuck to the surface of the
film. The films were then put on a glass slide using paraffin.
FIGS. 10-10B show SEM images of Bi filled films with 23 (FIGS.
10-10A) and 13 (FIG. 10B) nm channel diameters, respectively, after
partially dissolving the alumina matrix from the backside of the
film. The portion of nanowires without support tend to agglomerate
with each other, as shown in FIGS. 10-10B, and making it difficult
to resolve the individual wires. The fine Bi wires were amazingly
ductile (we observed from the SEM images that some wires were bent
by angles larger than 90.degree.) although the bulk Bi is very
brittle. This also indicated a good crystallinity of the ultra-fine
Bi wires. Since the whisker-like features shown in FIGS. 10-10B
were observed throughout the whole film, we concluded that the
porous film has been thoroughly filled by Bi.
The electronic states of a Bi nanowire array composites studied by
optical transmission spectroscopy. The porous alumina film, which
is clear and transparent originally, changed color after it was
filled with Bi, with a color that depends on the pore diameter of
the original template, i.e., depending on the diameter of the Bi
nanowire. Thus the 56 nm sample is dark, while the 23 nm sample is
yellowish and the 13 nm sample is essentially transparent. This
kind of color change indicates a dramatic change in the band gap
energy of the Bi nanowire as a function of wire diameter. This
color change is consistent with an estimate of about 45 nm for the
diameter at which the Bi nanowire makes a transition from a
semimetal to a 1D semiconductor. In order to determine the band gap
energy of the Bi nanowire composite, we studied its optical
transmission properties using an UV-Visible-IR spectrophotometer.
In carrying out the optical transmission experiment, the samples
were put between two glass slides. The spectrophotometer was
operated in the double beam mode and the incident beams were
perpendicular to the film.
Referring now to FIG. 11, the vacuum melting and pressure injection
apparatus presently used has several drawbacks which can be
improved for scale up production. The first drawback is the low
production rate. The second is the method of extracting the sample
from the solidified metal piece.
An apparatus 130 suitable for mass production is shown in FIG. 11.
In this apparatus space can be efficiently used inside the pressure
chamber by using movable multi-clamps 136, so that lots of films
can be filled during one injection process. The second advantage of
this design is that it bypasses the mechanical extraction process
by slowly withdrawing the samples from the liquid after the
injection is completed.
The operation steps of apparatus 130 are as follows. First samples
138 and metal pieces 140 are loaded into the pressure/vacuum
chamber 132. Next, the chamber 132 is evacuated to a desired level
using a vacuum means 144 and the chamber 132 is heated with a
heating means 134 to a suitable temperature to degas the porous
film 138.
Next the temperature of the chamber is increased to a point higher
than T.sub.m of the metal to generate molten metal. The samples are
then thoroughly immersed into the molten metal via clamps 136. Next
high pressure Ar gas is introduced into the chamber via means 142.
After injection is completed, the metal filled samples are slowly
withdrawn from the liquid 140.
Next the chamber is cooled at a predetermined rate to allow the
molten metal to solidify and crystallize inside the pores of the
substrate 138 and the chamber is then depressurized.
FIGS. 12 and 13 show XRD patterns for the Bi-filled films with
different wire diameters. All peaks shown on the XRD patterns are
very close to the peak positions of 3D bulk Bi, revealing that the
rhombohedral lattice structure of bulk Bi also occurred in the
small nanowires. The XRD peaks are very narrow and no peak
broadening was observed within the instrumental limit, which
indicates long range periodicity of the Bi lattice structure along
the wire length. In FIG. 12, for sample (a), with an average
diameter of 56 nm for the Bi wire, only three strong peaks ((202),
(110), (012)) are observed. For sample (b) and sample (c), with 23
and 13 nm diameter, respectively, only two peaks ((012) and (024))
are found, both belonging to the same lattice direction. We
therefore conclude that the individual Bi nanowires are essentially
single crystal and are all similarly oriented. The lattice
orientation along the wires in sample (b) and (c) are very close to
the bisectrix direction of the rhombohedral Bi structure, which is
a direction along which electrons have very high mobility. This
single crystal orientation of the array of the Bi nanowires will be
very important for many applications, such as study of the
transport properties of 1D Bi nanowires. In the XRD patterns, we
did not observe peaks normal to the wire, which should be very
broad due to the small diameters of the wires. The disappearance of
these peaks may be attributed to the lack of intensities or to
their large linewidth so that they are buried under the background
generated by the anodic alumina and the glass slide.
Another phenomena which we observed in the crystalline structure of
Bi nanowires is the metastable phase generated inside the small
nanochannels.
Shown in FIG. 13 are XRD patterns of the 13 nm sample before and
after an annealing treatment. In the as-prepared sample, we
observed a small peak- ((202)), and this small peak disappeared
after annealing the sample at 200.degree. C. for 8 hours under
flowing N.sub.2.
Shown in FIG. 14 are XRD patterns for the 56 nm sample. The
as-prepared sample has a peak at a value
2.theta..apprxeq.49.36.degree. which doesn't belong to bulk Bi, and
the intensity was reduced after annealing at 200.degree. C. for 8
hours. The metastable phase related to the 49.36.degree. peak
probably is a high pressure phase due to the high stress induced by
the lattice mismatch between the Bi nanowires and the anodic
alumina template. If this is true, this is believed to be the first
high pressure phase generated inside nanochannels.
If there is a lattice stretch, we should observe some peak shifts
relative to the 3D Bi peak positions. However, we did not observe
any peak position shift within the instrumental limit. Since all
the peaks observed are along the wire axis, and Bi--Bi interactions
along the wire should outweigh the Bi-template interactions, the
absence of peak position shift is not surprising. The lattice
stress at the directions normal to the wire axis should be much
higher than that along the wire axis, but unfortunately, we were
unable to observe the peaks normal to the wire axis in XRD
patterns.
The Bi nanowire/ceramic composite structures were also
characterized by SEM techniques. The SEM images of the topside
surface of the Bi filled film (see FIGS. 8-8A) showed that Bi had
entered all the channels of the film. No unfilled entrance nor
excess Bi particles or pieces were observed on the topside surface.
A very thin layer of Bi (thinner than 10 nm) was found to coat the
topside surface of the film. This very thin coating layer of Bi
would be advantageous for future applications involving transport
properties, since it may serve as an electric contact for the Bi
nanowires.
As for the backside of the film, after removing the alumina barrier
layer, SEM images (see FIGS. 9 and 9A) showed that there was a
short distance between the majority of Bi nanowires and the channel
entrances, although some channels were filled up to the entrance.
This empty portion of channels could be removed by mechanical
polishing or chemical etching.
In order to confirm that the whole porous film has been filled by
Bi nanowires, we developed a method to dissolve the anodic alumina
matrix and expose the Bi nanowires. Since the SEM images proved
that Bi had entered all channels from the topside of the film, we
partially dissolve the film from the backside. If we can show that
Bi nanowires have filled the channels close to the backside
surface, then there is no doubt that the channels were thoroughly
filled by Bi. This is exactly what we observed from the SEM images
(FIGS. 10-10B).
Shown in FIG. 15 is the optical absorbance of anodic alumina
templates and Bi nanowire array composites with different wire
diameters. The anodic alumina film prepared by a 20 wt % H.sub.2
SO.sub.4 solution (FIG. 15 line a) did not absorb photons until a
wavelength below 420 nm was reached, corresponding to a band gap
energy of approximately 3.4 eV. For the anodic alumina film
prepared by a 4 wt % H.sub.2 C.sub.2 O.sub.4 solution (FIG. 15 line
c), the absorption edge slightly shift to the right, corresponding
to a band gap energy slightly lower than 3.3 eV. As for the Bi
filled samples, the 56 nm diameter sample (FIG. 15e) had an
absorbance close to 1 throughout the whole spectrum (300-3000
nm).
The optical absorbance of the 56 nm sample, for wavelength between
2400 and 3000 nm, is shown in FIG. 16. The absorbance slightly
decreased for wavelength larger than 2800 nm (FIG. 16a), perhaps
indicating that the Bi nanowires with 56 nm average diameter
already became a narrow band gap semiconductor, with a band gap
smaller than 0.5 eV. The reason that the absorbance of the 56 nm
diameter sample is not even higher, i.e., much larger than 1, is
due to the photons that are transmitted through the alumina
matrix.
The 23 nm diameter sample (FIG. 15 line d) had an absorption edge
starting at a wavelength of around 1300 nm with a significant
increase at about 1000 nm, corresponding to a band gap energy
between 1.1 eV to 1.4 eV. The reason that the curve does not show a
sharp rising edge may be attributed to the distribution of the wire
diameters in this sample. For the 13 nm diameter sample (FIG. 15
line c), the absorption starts at around 900 nm with a sharp
increase at about 650 nm. This indicates a band gap energy between
1.5 eV to 2.1 eV. The distribution of wire diameter also explains
the absence of a sharp rising edge for the 13 nm diameter
sample.
A band gap energy of about 2 eV may be too large for Bi to be
explained solely by 2D quantum confinement in the nanowires. It may
be explained by the stress-induced high pressure phase of Bi
nanowires inside the nanotubes. It was known that the rhombohedral
lattice structure of 3D bulk Bi will be stretched more along the
trigonal direction when under hydrostatic pressure. The high stress
due to the lattice mismatch between the Bi and the anodic alumina
might help this kind of lattice stretch thereby enhancing the
semimetal to semiconductor transition. In XRD patterns of the Bi
nanowire composite, we observed one metastable peak which does not
belong to the 3D bulk Bi, and this might be another sign of the
high pressure phase. So we believe that the wide band gap of the
ultra-fine Bi nanowire composites is due to both a lattice stretch
and the quantum confinement effect in the two directions normal to
the wire axis.
Having described preferred embodiments of the invention, it will
now become apparent to one of ordinary skill in the art that other
embodiments incorporating their concepts may be used. It is felt
therefore that these embodiments should not be limited to disclosed
embodiments, but rather should be limited only by the spirit and
scope of the appended claims.
* * * * *